Disruptions: Data Without Context Tells a Misleading Story

Description: Several years ago, Google, aware of how many of us were sneezing and coughing, created a fancy equation on its Web site to figure out just how many people had influenza. The math works like this: people’s location + flu-related search queries on Google + some really smart algorithms = the number of people with the flu in the United States.

Source: NYTimes.com

Date: Feb 24, 2013


In today’s digitally connected world, data is everywhere: in our phones, search queries, friendships, dating profiles, cars, food, reading habits. Almost everything we touch is part of a larger data set. But the people and companies that interpret the data may fail to apply background and outside conditions to the numbers they capture.  “Data inherently has all of the foibles of being human,” said Mark Hansen, director of the David and Helen Gurley Brown Institute for Media Innovation at Columbia University. “Data is not a magic force in society; it’s an extension of us.”  Read Rest of Story 


 Questions for discussion:

1. How can Data without context be misleading?

2.  Will Big Data eliminate error, uncertainty, and risk?


38 thoughts on “Disruptions: Data Without Context Tells a Misleading Story

  1. J. Pointer

    Data without context is like a rumor. People will believe it, even when backed by no truth whatsoever. This is dangerous. Individuals need to research the credibility and merit of the source of information before they put all their faith and belief into and before relaying this info onto others, causing a snowball effect. For instance, can you imagine the things people would be believing, if they were told the Inquirer Magazine was 100% truth? Exactly. We all know not to believe such wild things as the Inquirer is not a credible source of information. That being said, you can receive information from a credible source for one topic such as health related concerns from your doctor.. but I would not necessarily take advice for how to fix my car from him as he is not credited in that field. That being said, because Google is not really a credible source for health related information as they use algorithms instead of facts, however the Center for disease control and prevention is a credible source as they use facts in relation to areas.

  2. Yuliya

    As the article discussed, data without the appropriate context can be easily misleading. Numbers and facts can be useful but it is crucially important to understand the background they were derived from. The context of information is very important to note when analyzing any type of statistics, and the information extracted from such data can be very misleading in nature if the whole picture is not evident. False apparent relationships may surface if not all variables are accounted for. Interpretation is key to determining true correlates. To use the example of influenza, an individual may be performing a search from personal symptoms or research purposes. Data sets can show relationships between anything, it is how they are collected and refined is what will make them accurate.

  3. Jarett French

    Data can be very misleading as demonstrated in the end of the article in regards to the stair and elevator use. Unless all circumstances revolving around the stats are known, the stats are useless. As well as what is emphasized by Google attempting to tract the flu in the United States. One can not get accurate data based on search results since many people search out of curiosity, or paranoia in regards to thinking they are sick when they are not. Researches can draw many false conclusions and even implement systems based on this false data. Many situations have probably occurred where such n experiment was conducted, conclusions drawn, and systems implemented in an effort to change a system and make something more efficient, and the in the end it is all for not because the data used was used out if context. Many situations have probably occurred where such n experiment was conducted, conclusions drawn, and systems implemented in an effort to change a system and make something more efficient, and the in the end it is all for not because the data used was used out if context. Big data or bigger data will help alleviate such issues and problems, however nothing can ever be perfect. And since nothing can ever be perfect you can never eliminate uncertainty and risk all together, although you may be able to for the most part.

  4. Edwin Owen

    As it’s presented in this article, data without context can be easily misleading. Indeed, it’s really difficult to extract information from data if you don’t have any idea about the context which has created these data. The last story of the article is a really good example of this fact. Without context, you will only be able to emit hypothesis and you are not able to know if you are right or not.
    However, I think that in the future, it could be possible to use data without the context of their creations. Indeed, the technology is improving really fast so maybe it will be possible in some years to create algorithms which take so many variables into account that they will allow to make conclusion based on this kind of data. However, the example of Google shows us that creating successful algorithms is really difficult because you have to put into a formula all the components of the real world (which is really complex and changing). However I stay confident in the fact that the technology will be able to overcome a majority of issues we are facing right now, including this one.

  5. Jingyi Wang

    Data without context can be very misleading, because the researchers only focus on the data and ignore the outside world that will cause an effect on the data will finally make a wrong or inaccurate conclusion. For example in the New York University experiment, they prepared a lot for collecting data and found students use the elevators in the morniong while walk stairs in the evening, and they give their conclusion that students are tired from sleep in the morning so they use the elevator and they are energized in the evening so they choose to walk stairs. This result seems quite reasonable and the researchers are satisfied with their conclusion until one day the security told him that the elevators are broken those days, students had no choice but to wal stairs. Actually they choose elevators no matter which peirod in the day. As we can see in this case, the researchers were mislead by the data, they didn’t pay attention to the context of it.
    Big data can help eliminating errors, uncertainty and risks in some circumstances, but we should still be aware of the context and background of those data. As we found in the Google equation. This case warns me that before making conclusion, think twice why this happen, not just make up an excuse but do make sure the causes.

  6. Adam Christensen

    Data context can be misleading because sometimes you could search for one thing thinking you have the correct “keywords” and get something completely different then what you expected to get. For instance, in this example people could’ve been searching the flu for topics on how to prevent it rather than treating it. This would cause flaws in the system and the statistics that are outputs due to the searches.
    Big data will help lower the chance of error, uncertainty, and risk because there we’ll be more data to get statistics from but the same issue will remain with the context. Another issue that will remain is the challenge of creating algorithms to withstand the amount of data that there will be.

  7. Warren Murley

    When conducting research it is important to be open minded, ask the question “Why?” and to not jump to conclusions.
    Data without context can be very misleading because, like most things in this world, when things are taken out of context error usually follows close behind. Data with out context is just numbers or words (or what ever was collected) and is then left to be assumed that the experiment worked or it didn’t work, or that the data is saying “this” but in reality it is saying “that”. There is a bigger picture out there and the collection of data doesn’t capture that whole picture. It generally captures a portion of that bigger picture and that portion usually contains errors.
    I don’t believe Big Data will eliminate error, uncertainty, or risk. However, I think it can definitely reduce the error, uncertainty or risk that comes with collecting data. It is important to remember the bigger picture. Data collected from verbal conversation, observation, or by request (surveys) will always contain some error,uncertainty or risk. I believe that it is the researcher’s responsibility to do what he/she can to reduce error, etc. and they do that by considering other factors than the data itself.

  8. Callie Matz

    Data without context can be misleading because it does not look for other reasons that a result occurred. In the University example they ignored the elevator being broken. Had the elevator been working they would have had a very different finding. In order to avoid skewing results, controls need to be in place in order to ensure consistency within the experiment. This is evident in many other fields of study such as psychology and science where they create a control group and manipulation group so they can base their results on reliable measurements.
    I think there will always be error, uncertainty, and risk when using big data. There will always be variables that cannot be controlled and some element of the context that cannot be understood. I think that it is important to look at data with a skeptical eye whether you are the researcher producing the data or someone searching for information. We must be sure that we look into who is funding the research and the methods used to obtain the information. I think that companies like Google should be more transparent when presenting big data by indicating how they came up with their results and expressing context behind the information.

  9. Kelsey Allan

    Data without any context can be very misleading as we have seen in this article in goodle’s algorithm and NYU Experiment. The issue with having no human involvement within both of these studies was evident because little things that affected the experiment were not picked up on by the computer. This means that we cannot rely solely on computers to turn or raw data into interpreted data. There needs to be some sort of human involvement that would allow for things that could have mislead the experiment. Like for google’s example, the search needed to be specific, there could just be people looking up what it is after the newscast or looking it up to question if they had it. However, these people may not have had the flu. In the NYU experiment, the context needed to be put into their study, if they had some human involvement in their study they would have found right away that their data was not useful because the elevator was broken one day and students had no choice but to use the stairs.

    Using “Big Data” can help to eliminate some of the risks that we have seen in both google’s and NYU’s experiment. However, if there is no context to the study and no human involvement within the study, ultimately the data that is interpreted will most likely be misleading. No matter how much data is collected there needs to be a context without one, we are just lost looking at misleading data everyday.

  10. Matthew Chipman

    Data without any context can be misleading as well as easy to manipulate. First off, if you take data in its raw form it can be manipulated to support really anything. i think it is important when collecting data, that the data collected is used for a specific purpose. Alot of data collected from surveys and market research is stored in databases and used to compile statistics and make conclusions based on data that was not collected for that purpose. This often result is inaccurate conclusions being made.

    Big data may provide some benefits to data collection and observations. It may weed out a few of the inaccuracies and misleading data that is collected. HOwever it is never 100% certain that you can project the future and determine with absolute certainty that a statistic proves a particular outcome. There should always be strict permaeters in place and several studies done and studied carefully in order to provide the most accurate hypothesis possible. Data is in all the world around us, alot of it is true alot of it was made up by millions of people and we havee to be careful about what we believe to be true and accuarate.

  11. Alicia Dyck

    Data is an important part of learning. Being able to get the appropriate data for any experiment or learning opportunity is equally important. Without accurate data, the conclusions that are drawn cannot be correct or taken to have value. In order to record correct data, it is important to understand the context in which it was taken. For this reason, if a computer is recording data, like the students taking the stairs or elevator, it would be necessary to ensure that the people analyzing the results understand the context. This is because the results rely on the context. This is why, in many science experiments, the experimenter will set up a control group and ensure that any differences between the control and other experimental groups are recorded accurately and effectively. Without knowing the exact context of the data, results cannot be derived correctly.
    If the data that is recorded is out of context, or the context is not explained to the people who will be utilizing the results, it could be misleading. The reason for this is because without knowing the context, many people would choose their own context for how the data was collected and choose to analyze in the specifics that they themselves have decided upon. This could be potentially very dangerous because the individual using the data may choose the wrong context and use it in a way that is incorrect or detrimental to others.

  12. Litchi Peng

    Data without context can be misleading. According to different cases, we may come up with different solutions. People do the research to get the data so that they wanna get some ideas to see if there is a rule or regular pattern in some specific things. If we get rid of the context to analysis data, it is useless. People may get the wrong information. In this case, people get an equitation via computer. The results may be very general or nonsense. If goggle try to get the data with our consider all of the other aspects, they will not can make the right interpretations. I think that Big Data will reduce error, uncertainty and risk. However, it still can not get prefect data for sure. Computers can gather different data from different fields, that will influence the accuracy of data in the context. That means the more information computers get, the more accuracy results we get. However, computers are not perfect, they may have mistakes as well. I do not believe computers can get the exactly results with so many variables. It had better people collect information, and only use computers do some simply analysis like equation analysis.

  13. Jarrett Potvin

    Data without any context can be very misleading. For one if you take data in its raw form it can be manipulated to make something look better or worse. The example in the reading of how this can happen is Google’s over estimation of influenza cases in the US. This happened because Google only looked for the word “flu” to be searched and not the context it was originally in. Therefore there was no discrimination between those looking to self-diagnosis themselves with the flu and those looking up how to best protect themselves from contracting the illness. It is those simple oversights which pulls data out of context that can make data misleading. Big data will not remove error, uncertainty or risk from the interpretation. Sometimes more data is not always better but in factor aids in the data being presented to be more misleading. Even with all the data that any researcher could want at his disposal judgment errors as well as biases still occur. While machines cannot be biased towards information the programmers and interpreters can and will be. Also the assumptions made can influence the data. Take for example from the article with the study about the elevator and the stairs. The major assumption is that the data was accurate but failed to take into result one of the elevators was malfunctioning. This just means that data itself cannot take the error out interpretation.

  14. Josh Bodnaruk

    Data without context can be harmful because it can be used to misrepresent the actual situation at hand. Data that suggest relationships especially needs to be considered “in context” just to simply make the correct assertions. I believe that there is a great potential for the ability to collect information from a large population. But there is also inherent risk, especially if the information was obtained by the wrong sources. However, the ability for almost real-time information allows everyone to benefit.
    Unfortunately, no matter how fine and precise information is collected, there is never going to be absolute certainty. There are constantly factors changing in the world that no single person or entity has direct control of. Furthermore, even if data is collected on an instantaneous basis, the data is already data, for it absolutely impossible to know future events. Therefore, any prediction will contain risk.
    In the specific case of medical data however, I believe that being able to give any advice is better than none. The ability to consult with physicians about various conditions and illnesses is a tremendous thing, and it’s something we almost have here at home. Were it not for the laboratories that process all the blood and other things for tests, we as people would have a substantially lowered quality of life. Therefore, I believe that it is absolutely necessary to be constantly tracking medical information. Only we have to hope that such information could not get into the wrong hands and be used for horrendous things.

  15. Chance Olsen

    Data without context can be very misleading as we read about the NYU experiment. They had come to the conclusion from data gathered that students used the elevator earlier in the day and then used the stairs later at night, this in fact was wrong since the elevator broke and people had no other choice but to use the stairs. The computers that were gathering the data had no idea about the external factors that were in place, as is the problem in many instances when data is collected. When Google tried to make an algorithm for people in the united states who had the flu, they would have had to take into account the people that searched the flu for many different reasons such as, experimental reasons, curiosity, and people that actually had the flu and were looking for cures. They tried to make an algorithm that would take into account many of the external factors that were in play already but when their algorithm says 11% of the U.S.A had the flu in January which doubled the Centers for Disease Control and Prevention estimate of only 6%. The large difference is due to a ton of external factors that could not be calculated by Googles algorithm.

  16. Alphine Bindiridza

    Data without context is quite misleading because one does not have enough evidence to support the data that he/ she found. People need to go deep into the story when they get data. they need to find out the causes, implications, and effects of the data. That way they will be able to trust the data that its valid and be confident to share it with other people. Posting numbers without the context is not good at all because when people find the correct information and discover the figures posted were misinterpreted, they lose trust on those websites with the wrong figures. Computers can collect data but can we reply on them? I think to some extent humans have to do some of the work because computers are not able to do most of the stuff especially those that have to be done manually. Big data can not eliminate error completely but it will help people get close to the correct information needed not just coming up with figures or percentages without context.

  17. Stacey Demchenko

    This is a perfect example of how data can be misleading, and this is a mild example. When there are only a certain number of factors involved in data collection, interpretation, and follow up-it is impossible for precise, valid and completely accurate conclusions to be drawn. As a few of the others have stated, and I will have to agree..almost all of the information you hear needs to be taken with a grain of salt. If you talk to any statistician, they will say that ‘correlation does not imply causation’. If you ever get a chance to look into the example of airbags reducing death in car accidents-you will see that the data collected has been skewed and calculated in such a way that it seems as though airbags are helpful. When you dig deeper, and dissect the data (with the help of a statistician, as it can be a trying task), you see that actually the airbags are doing more harm than good. I do believe that Big Data could reduce the probability of uncertainty and risk, but even then complete accuracy will not likely be achieved. There are many factors, both internal and external, that a computer can not account for. Until all variables, factors and influences can be accounted for when interpreting data, we should not have 100% confidence in the results.

  18. Leisha Hansen

    When interpreting data without context, researchers could find ways to prove anything they wanted. Without having controls, checks and balances set up with the experiment there is no way to determine what the data actually proves. Like in the article, by simply using data and no observation, the researchers had no idea that a big factor in their study was compromised (the elevator being broken). This experiment was uncontrolled which lead to a great error in the interpretation of the results. In uncontrolled experiments there could be many factors influencing results that need to be recorded and brought into the analysis of the data in order to achieve more accurate and valid results.
    Big Data predictions will always contain error, uncertainty and risk. Attempting to predict the future is rarely 100% successful because there are so many uncontrollable influencers on its outcome. Data can help with planning for the future, but it will not tell you that tomorrow there will be a massive fire that breaks out and destroys half a city. The events that the future holds are unpredictable and their influence on the predictions made are highly uncertain. It is always a risk to make decisions based on predictions, one just has to hope that they choose the best possible alternative based on the best knowledge about the future that they can obtain.

  19. Sheri Durina

    Data without context can be misleading in all kinds of situations; from the collection of big data to overheard information from a conversation. Data is collected for many purposes and can be very useful, but without the appropriate dissection and interpretation, only part of the story is being relayed to the end user. On a smaller scale this can be comparable to the small pieces of information that can be overheard from a conversation.

    Collecting big data can be very useful; however, without all of the information or without looking at the “big picture” the data can cause inaccurate or simply false results. When collecting this data, researchers may have the best of intentions, but without interpreting the data so it is useful and accurate, there can be confusion, misunderstandings, and in the case of the Google influenza experiment, wide-spread fear.

    As mentioned earlier, data being taken “out of context” can be related to portions of overheard conversations. A piece of data taken “out of context” can spread very quickly from person to person and soon a potentially false rumor has been spread to many. Data taken out of context can occur on a large scale or in much smaller situations, either way the story is often misleading.

  20. Mariam akinola

    A big problem with data is the lack of context. Data without context can be very misleading because it doesn’t allow the user of the data to understand and grasp the whole idea and what the data is all about. The computer doesn’t ave the ability to know what human beings think and it can’t produce anything which hasn’t been entered into the computer. Other factors can affect the data which can make is complex and misleading to any user of the data. The experiment performed by the students in NYU was not accurate because the elevator broke down and people used the stairs instead. On a normal day, most students will prefer to use up the elevator instead of the stairs. Also, the google algorithm results were double the actual estimate for influenza. This was because the google search considered all other kinds of flu not just the influenza.
    Big data can also create a major problem because there are so mangy dates all round us and we don’t know which one to use or which ones are accurate. We find data on our phone, social networks, offices, everywhere. It cannot eliminate uncertainty because there are so many possible answers to what you are looking for and you might not know which one is valid or not. The risk can be greater if we don’t carefully select the right data to use. Lots of companys pay a huge sum of money to be on the first page of any search engine for people to look at. So whatever kind of data you search for on your computer will generate millions of possible answers to your question and you have to be very careful in selecting the right kind of data.

  21. Brad Zander

    This article was very interesting. The examples bring to mind the fact that one must:
    1) Always look at the big picture – One must look beyond the data that is received. There is always a story behind the data. This story will help us make correct decisions and allow us to avoid making wrong decisions by misrepresented data.
    2) Always ask why – It is important to ask why. The article speaks clearly to this necessity with the example of the elevators. The flaw in the data was revealed when the security guard shared the why as to why the students were actually using the stairs more. This idea applies to almost every aspect of life. Many things happen and we and others choose to do many things. It is important to understand why we and others do the things that we do. This will allow us to make correct decisions.
    Companies, such as Google, need to ensure that they make full disclosure of the purpose and context of the things they do. This will protect them from overly critical people and protect them from scorn that could come from outsiders. This is shown through the example found in this article.

  22. Kevin Robichaud

    I believe that data without context can be extremely misleading. The reasons that alot of people google flu symptoms is because they are inquiring to if they have it not confirming an assumption. Also when people start to hear Numbers thrown around of the amount of people with the flu being high it will spark intrigue from people who are not educated on the sickness. This intrigue will lead them to the easiest source of answers, a google search which i believe would inflate the number of potential virus sufferers and create a falsified report on the number of people with the flu. Also The amount of repeat searches on a certain topic could cause a misleading number of sufferers. The same could be said for it being understated. The amount of the population that uses different or no search engine could potentially be huge, especially with older victims who are not tech savvy. In my opinion there are to many variables that are not being accounted for. The fact that google is willing to back these numbers without acknowledging the other search engines is of major concern. As a rough estimate i could see using the data but anything that requires a certain level of accuracy makes me question the numbers. For example if i was a major producer of medicines i do not think i would use these numbers as a means of deciding how much medicine should be produced.

  23. Chyane Tibert

    The problem with data is that it can be interpreted differently giving different results depending on the context of the situation. A computer does not take certain variables into account and can only calculate data the way you program it to. This possess as a problem because we are relaying on computers to interpret information more and more without questioning the validity of the results. We also need to take into account weather the results are within a reasonable range that we would expect and compare to other results, such as pervious years or other sources; we should not just accept data as absolute.
    Big Data will reduce error, uncertainty, and risk but we also need to accept the fact that it cannot absolutely eliminate these issues. The more data that is collected the more accurate the results will be, variances such as the elevator breaking down for a day in data collected over the week will have a large affect, where as if the data is collected over a month or year a day will have little to no effect on the results. Collecting Big Data will also causes the problem of greatly increased expenses and the time to collect the data.

  24. Yu Chao

    Data is not usually correct and it can caused the misleading . data is the tool to collect people activity and analysis the tendency of people’s behavior . But people and their behavor are always changed, and it is hardly to use the right data to show the real life

  25. Megan Jackson

    Data without context can be misleading in that without knowing where the numbers came from you do not have enough evidence to make a conclusion on the data. Without know the variables that affect the data that a study collects making a conclusion is meaningless. As with the case in the article where Google had predicted the amount of people with the flu with out taking into consideration the variable that could have affected their results such as the widespread media coverage on the flu virus this year.
    I don’t think big data is going to eliminate error, uncertainty and risk because with all data there is always a chance that there may be an error. You can never be 100% on something.
    I think with all this ‘big data’ out there data will not become easier to predict but instead harder to predict. There is so much data out here now with the Internet that the big data is just to large for us to include every single variable. Like with what Google did: there was just too much data and variable that they couldn’t have possibly included everything and like with the students who tried to predict if students used the elevator or the stairs more. They could not have predicted that the elevator being broken was what led them to the results that they got. However, once we learn how to apply context to data numbers big data is going to be very powerful.

  26. Michael Jensen

    I agree that Data can definitely provide estimates, however, data does not take into account every single other context in which people engage in. In other words, the data is rather limited to only a small number of contexts. Subsequently, this means that many conclusions will not be entirely accurate (i.e. Google’s equation to calculate how many people have the flu). The world is constantly experiencing change, and it is hard to keep up with all of that change, and is a key component in calculating accurate information about a subject.

    I agree that big data will eliminate error, uncertainty, and risk, but only to a certain extent. With regards to big data, research is far more broad and inclined to gather more sustainable and accurate information, but not 100% perfectly. There will always be a lack of data that will not be accounted for, which will not give a lead to entire accuracy. It would be far too costly to obtain information for every single context that an individual happens to be participate in. That being said, I believe that it is a breakthrough and a great effort in predicting information that could potentially help protect and educate the public in everyday life.

  27. Lanre Paulissen

    This is the age of big data and for the most part I think it has come to stay and harnessing its potential would keep evolving in the coming years.

    Data is useful but when all conditions, preconditions, and or even post conditions, which might be difficult to ascertain, are not factored in, it can be misleading because the data would definitely be out of context and as a result the analysis of such data might result in under or over statements as in the case of the Google Flu Algorithm. The elevator vs. stairwell use at the New York University is another classic example. The conclusion of those who carried out the survey was much skewed, “students seemed to use the elevators in the morning, perhaps because they were tired from staying up late, and switch to the stairs at night, when they became energized.” The experimenters should be thankful to the security personnel who realized the bias in the analysis.

    I incline to the school of thought that big data is able to help reduce errors, uncertainty, and risk but it would definitely not eliminate all the errors. Besides, it’s not just about the volume of data; it’s also about the quality of the data. The quality is what I believe is more important as the analysis of such data would go a long way in being able to have confidence in the result.

  28. Lauren Gallimore

    Data, when used without context or in an inappropriate context can lead to inaccurate conclusions. Computer programs and algorithms are only able to capture so much of what influences the data. The amount of data and information a computer is able to analyze is huge, however, it is impossible for the computer to come up with perfect information. The world is always in a state of change and there is always new information, so it is impossible to be 100% correct all the time. The data going in to be analyzed may have come from unreliable sources like the information used by Google Flu Trends. It is impossible to validate these sources and ensure accurate conclusions are drawn. When Big Data takes more data into consideration, it decreases the effect of statistical anomalies and incorrect sources on the overall conclusion formed by the data. The scope of data that computers are able to analyze is much more than any group of scientists would be able to do. This is why we rely so heavily on computers and the information they give us. Although computers can often give us useful information, people must be critical of the results and not blindly trust that it is always correct.

  29. Brennan Lowe

    I think it is safe to assume that big data will reduce error, uncertainty, and risk drastically. I do not believe, however, that big data can completely remove error, uncertainty, or risks. In a few years, when our technology continues to expand and become smarter, will remove all errors, all uncertainties, and all possible risks. Today, computers have great big data capabilities, and can be very accurate given the proper algorithms and programs. You look at this article, and big data could not properly calculate the amount of American’s with the flu. Google’s algorithm doubled the percentage of actual Americans getting sick. However, referring to the student seeing if people use the stairs or elevator more, big data can easily compute something in this nature. It can calculate something easily with no errors in a day task, but in order to use big data for one of the most populous countries in the world may need to be a little more advanced in order to achieve the goal.

  30. KJ

    Data without context is like having knowledge without application. Take for example, learning a martial art from a book – it’s not the same as going through the application of doing the techniques and conditioning the mind, body, spirit experience. Data without context is just that – without someone with the ability to clearly interpret what is being said; it just sits stagnant. Have the data is good, but knowing how to apply it is the most important thing. Will it eliminate error/uncertainty and risk to have the information? Not on its own, but the application of that information may if it is correctly applied to the circumstance.

  31. Carla Hornecker

    This article is a clear indication of how data without context can be misleading. As humans, our interactions, ideas, and communications are so complex and dynamic that individual pieces of data cannot always be taken at face value. Every piece of data exists within an external environment with many other potential influences that affect its meaning and intended interpretation. If computers try to “understand” this data without taking into account all of the other influences affecting and forming it, they will not always be able to make the correct interpretation, as illustrated in the case of Google’s algorithms.

    I think that Big Data will certainly reduce error, uncertainty and risk, but I don’t think that there’s much chance of it eliminating these things altogether. Definitely the more information computers have, the more factors they will be taking into account when interpreting data, which will increase their accuracy. However, so many variables and influences make up the context of data that I don’t think that computers will ever be able to ensure complete accuracy all of the time. I think that it is, and always will be, wise to take any information given to us by computers with a grain of salt.

  32. Devin Phalen

    Data can be useful, useless, or used to promote an agenda. For instance, as contradictory as it seems–two individuals can take the same statistic and both use it promote their own differing agendas. Politicians do it all the time! Unemployment levels dropping from 8.0% to 7.6% are promoted as a positive indication by the party in power (knowing the general population doesn’t know any better). The opposition, on the other hand, knows that the rate only includes those looking for work, so discouraged job seekers dropping out of the job hunt could also lead to unemployment levels dropping. Correlation doesn’t imply causation, and data isn’t always perfect. For these reasons, I’m pretty leery about taking someone else’s data at face value. I feel like you need to use your own intuitions and judgement about the legitimacy of the source of the data. That being said, I do think that taking in a broader and broader array of data does help reduce risk and uncertainty. Just like diversification of assets in the financial world is a hedge against losses, diversification of sources in the intellectual world is a hedge against misinformation.

  33. Jacqueline Wegener

    Yes data can be useful in collecting information but it may be misleading in the sense that it does not capture the entire story. There are many external factors that can skew the results that data presents to us. In the example of the elevator at N.Y.U the students conducting the experiment believed that they had come to the conclusion that elevator use increased as the day progressed. This however was not the case as the elevator broke down and the students were forced to use the stairs which caused a flaw in the data that was presented.
    Computers are not able to be aware of external factors and do not take these factors into account when collecting data. Also, it would be more beneficial to use a larger time frame for collecting data as this will decrease the effect that random external factors can have on the results. If the data regarding elevator vs. stair usage was collected over several years the results would have been more accurate. Testing elevator usage over a shorter time frame and having an elevator break down during this time causes many discrepancies because if it were only being tested for 2 days then half of the time that you were collecting data students would have no other choice but to use the elevator.

  34. swatigade

    As we saw in this article, Data without context can be misleading as it does not consider all of the factors involved. Data can be of different sorts and it is up to us to interpret it based on our needs. In this article, the computer only measured the sensors such as activity/movements of the students. It is up to us to interpret the results of the computer as well as take into consideration of all other known factors. If we just rely on the data then the results can be misleading. Big data can eliminate error, uncertainty and risk but again you cannot rely solely on the computer output. You have to consider others factors as well. The larger the data the more information we have and if something does go wrong at a certain point, by having large data it will not have a big impact of the results. So in the case of this article, if the research was done over a longer period of time then having a broken escalator for a couple of days would not have affected the results as much. Data does not cover all the facts and one must be cautious in using and interpreting it.

  35. Akina Morimoto

    Indeed, data is valuable. However, it is only valuable if it used in the right context with the right information. Since this data is collected based on an algorithm, there are going to be other factors that affect this and make the data irrelevant or incorrect. Computers are valuable and useful, but there are some elements that computers sometimes cannot comprehend or take into consideration. Using computers to collect data can be efficient if the likelihood of other variables being a factor is limited. However, it is not always possible to predict the outcomes of an experiment or study. Big data certainly could eliminate error, uncertainty, and risk, but only to a certain extent. When there is a larger sample size of data, it could eliminate the errors that the computer may present. However, if this is the only way for computers to generate accurate information, then this is likely not an efficient and economical approach to collecting data. This would mean that it would require more time to collect more data, as well. In this case, it would make more sense for humans to collect this data and organize it, then use the computer to finalize and interpret it.

  36. Richelle Merrick

    This article perfectly points out why data without context can be misleading. The study about the elavators and stairs was the best example on why you need an entire story to give proper information. Computers are a great source of information but to collect the data, humans are a necessity. A computer can capture numbers but not context. There are circumstances in every situation that need to be added in to the equation and without including this your data is wrong. For example, People want to know what time the average person eats supper. A computer can tell you how many people ate at a certain time during the study but it doesn’t take the circumstances into account. Say a couple of the people involved had appointments, late days at work, or a special celebration that were a change in routine. A computer is not going to include this. It comes down to the fact that you cannot rely solely on what a computer tells you. All information is not 100% accurate no matter what the source. When you get data off the computer just make sure to not believe it is an exact reality. There is always going to be error, uncertainty, and risk. These things are inevitable. However, with technology advancing as fast as it is there will be a huge decrease in these things.

  37. Jonny Kostiuk

    Data without context can be very misleading because we do not know what outside factors affect the data, such as the elevator breaking down. Data will never be true unless all the facts that could affect the data are know and manipulated in a way that prevents them from disrupting the data process. Computer can not always detect all on these manipulative outside variables that humans can and that is why the data would be misleading. Big data could help in reducing in error, uncertainly, and risk but it won’t eliminate those problems in this day and age. When you have big data you have extremely large amounts of data that need computers to process is because of it size and the more data you have the more likely error is reduced. If you tested the elevator and stairs for three days and the elevator broke down for one night your results would be pretty misleading but if you recorded that data for 1 year and the elevator only broke down a few times it would not make a huge difference in the data collection.

  38. TJ Winn

    This is a classic example of needing to take all information, especially these days, with a grain of salt. Though I do think data can be (and already is) extremely useful it is all in the way the data is interpreted and by who. Computers have no problem collecting, storing and regurgitating data… I am sure they can even interpret it. However, just as we read, it is obvious that we can’t (not yet anyway) rely solely on computers to do all the tasks of humans. There are human and other ‘variables’ to each experiment that computers cannot always detect. Furthermore, there is context of words and situations that is also, most likely, undetectable by computers. This is how the data is taken ‘out of context’ and thus the data and conclusions drawn from that data are invalid.

    I think that big data can mitigate error, uncertainty, and risk and also provide priceless insight to problems and issues; however I don’t see, in the immediate future, that it will eliminate these things. Thus we humans need to remember that data is just data and that computers, though labeled ‘smart’, are simply doing what we command them.

    In short we need to remember the trite, banal saying of GIGO; which stands for garbage in, garbage out, no garbage in gospel out.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s